This paper applies the concept of the applica- tion server, three-tier architecture, EJB, and J2EETM to knowledge discovery in large databases which is using the letter recognition classification and detection as an exam- ple case. EJBs will be used as intermediate connecting layer, also called middle tier, to get requests from the clients and pass the request to database server. Cloud- scape or Oracle9i database server will provide database support and JDBC (thin) for cloudscape or Oracle9i that work on Windows NT/2000/XP environment will provide database connection. We show how a classification metric drawn from Rule-InductiordDecision Tree algorithms can be computed via the Java model. Performance evaluation figures, such as classijication error rate and number of rules, are presented for adaptive data mining classifiers and jlters, using the Letter Image Recognition Data ex- ample which is the practical implementation of this work.
[1]
Ian Witten,et al.
Data Mining
,
2000
.
[2]
Philip K. Chan,et al.
Advances in Distributed and Parallel Knowledge Discovery
,
2000
.
[3]
Moustafa Ghanem,et al.
Large Scale Data Mining: Challenges and Responses
,
1997,
KDD.
[4]
David J. Slate,et al.
Letter Recognition Using Holland-Style Adaptive Classifiers
,
1991,
Machine Learning.
[5]
Alberto Maria Segre,et al.
Programs for Machine Learning
,
1994
.
[6]
Kathleen McCABE,et al.
On-Line Documentation
,
1997
.
[7]
Steven L. Salzberg.
On Comparing Classifiers: A Critique of Current Research and Methods
,
1999
.
[8]
Karsten P. Ulland,et al.
Vii. References
,
2022
.
[9]
Ian H. Witten,et al.
Data mining: practical machine learning tools and techniques with Java implementations
,
2002,
SGMD.